Vision based matching for landmarks has been widely studied and applied in the field of visual search. For example, for an indoor venue, images around an indoor venue can be collected. For each image, scale-invariant feature transform (SIFT) features can be computed and stored in a database (DB). Each image in the DB has associated location information. This information can be mapped to a location or an area on the venue map. A query image captured in the venue can be compared again images in DB based on feature matching. Its location can therefore be determined and mapped. However, conventional methods have not taken advantage of newer devices that are equipped with dual cameras.
Therefore, there is a need for system and method of determining location of a device using opposing cameras.